Cooperative Bat-Algorithm-Based Localization and Tracking in Networked Multi-Agent Systems with Noisy Range-Only Measurements

Abstract

Networked multi-agent systems often require accurate localization and tracking using limited sensing and bandwidth, and range-only measurements are a common choice due to their low cost and energy efficiency. However, range-only sensing introduces nonconvexity, unobservability in certain configurations, and sensitivity to noise and outliers. Cooperative strategies that exploit inter-agent communication and motion coordination can mitigate these limitations by sharing information that improves identifiability and robustness. This study investigates a cooperative approach to localization and target tracking in mobile agent networks in which agents possess only noisy inter-agent and agent-to-target ranges and rely on motion primitives that can be actuated with bounded accelerations. The proposed method embeds a population-based global search mechanism within a consensus-driven estimation layer to reconcile global exploration with local statistical efficiency. The cooperative layer integrates local measurements and neighbor messages over a time-varying interaction graph while maintaining mild requirements on connectivity and asynchrony. Performance is assessed with respect to estimation accuracy, convergence behavior, computational scalability, and communication load using a set of representative trajectories and network topologies. The analysis emphasizes the interplay between geometric observability, excitation of agent trajectories, and adaptive parameter selection that jointly determine practical accuracy. Numerical experiments illustrate error reduction at moderate signal-to-noise ratios and improved resilience under sporadic link failures. The approach is presented with implementation considerations that reflect realistic constraints on sensing rates, latency, and clock offsets. The results indicate that a neutral balance between heuristic exploration and principled estimation can yield consistent localization and tracking performance under range-only sensing.